Text analysis and data storytelling project applying topic modelling and sentiment analysis to 295 public submissions to the 2025 Australian federal election inquiry. Graded H with 90/100.
An exploratory text analysis project applied to 295 public submissions to the Australian Parliamentary inquiry into the conduct of the 2025 federal election, published by the Joint Standing Committee on Electoral Matters (JSCEM). The inquiry question explored how submissions characterise the relationship between digital misinformation and electoral intimidation, and how that perception varies across submitter types. The pipeline uses NLTK, spaCy, Gensim LDA, VADER, TextBlob, and scikit-learn TF-IDF for topic modelling, sentiment analysis, and submitter classification. PDF extraction handles both native-text and scanned documents via pdfplumber, PyMuPDF, and pytesseract with an OCR fallback. Findings were synthesised into a cohesive markdown narrative report combining visualisations and written analysis. Completed as part of the Applied Natural Language Processing subject at UTS. Graded H with 90/100.
